There has been little study of occupational mortality in Hispanic populations, despite evidence that Hispanics have greater potential for hazardous exposures in some occupations, and evidence that Hispanics have unusually high rates for certain diseases. For example, tumors of the gallbladder, stomach, cervix, and liver occur excessively in the Hispanic population of the U.S., but the etiology is unknown. A major reason for lack of study of occupational mortality in Hispanics has been difficulty in calculation of death rates. The numerators (Hispanic deaths) and denominators (census estimates of Hispanics by occupation) have not been comparable because of differing methods in ascertainment of Hispanic ethnicity. The 1980 census, for example, used self-reporting of Hispanic ethnicity to estimate Hispanic populations, a method that cannot be used for deaths. We propose to calculate Hispanic mortality rates for occupations in California using an innovative method of controlling for bias caused by differing methods of Hispanic ethnicity determination. specific aims are as follows: 1) ?Calculate Spanish surname/self-reported Hispanic origin comparability for each occupation, using the U.S. Census 5% microdata sample for California that contains Spanish surname, Hispanic origin, and occupation, 2) Compute Spanish surname population estimates for occupations in California using the above comparability calculations and the U.S. Census 20% sample for California that contains occupation and self-reported Hispanic origin, 3) Select Spanish surname deaths from the 180,000 deaths in the California Occupational Mortality Study in which occupations were coded for all state death certificates for a 3-year period, 4) Calculate age-adjusted mortality ratios for each occupation for the 92 standard NIOSH causes of death categories using mortality rates in all occupations combined as the comparison, and 5) Compare mortality rates for Hispanics and non-Hispanics in each occupation. Since the study will examine many occupations and causes of death, and will not account for smoking and alcohol, it will be considered a hypothesis generating study. It may provide direction for future cohort and registry-based case-control studies where it is possible to focus on specific illnesses and injuries, exposures, and confounding factors.